- Email: [email protected]
Postsplenectomy course in homozygous sickle cell disease
P
Postsplenectomy course in homozygous sickle cell disease
J. G. Wright, MB, MRCP, MRCPath, I. R. Hambleton, BA, MSc, P. W. Thomas, PhD, N. D. Duncan, DM, S. Venugopal, MS, FACS, FICS, and G. R. Serjeant, MD, FRCP
Objective: To determine whether children with homozygous sickle cell (SS) disease and splenectomy are at greater risk of death, overwhelming septicemia, or other complications. Methods: A total of 130 patients with SS treated by splenectomy (46 recurrent acute splenic sequestration, 84 chronic hypersplenism) over a 22.5-year period at the Sickle Cell Clinic of the University Hospital of the West Indies, Kingston, Jamaica, were compared with a control group matched for sex, age, and duration of follow-up in a retrospective review. Deaths and bacteremias were examined over the whole study period. Painful crises, acute chest syndromes, and febrile episodes were compared in the 90 patients completing 5 years of postsplenectomy follow-up. Findings: Mortality and bacteremic episodes did not differ between the splenectomy and control groups. Painful crises were more common in the splenectomy group than in the control group (P = .01) but did not differ between splenectomy indications. Acute chest syndrome was more common in the splenectomy group than in the control group (P < .01) and was more common in the acute splenic sequestration group than in the hypersplenism group (P = .01). Febrile events did not differ between the groups or between the indications for splenectomy. Conclusion: Splenectomy does not increase the risk of death or bacteremic illness in patients with SS disease and, if otherwise indicated, should not be deferred for these reasons. (J Pediatr 1999;134:304-9)
The complex vasculature of the spleen provides an efficient filter against bacteria, and the mass of reticuloendothelial cells in close proximity to the From the MRC Laboratories (Jamaica) and the Division of Paediatric Surgery, University of the West Indies, Kingston, Jamaica.
Submitted for publication May 8, 1998; revision received Sept 18, 1998; accepted Oct 28, 1998. Reprint requests: GR Serjeant, MRC Laboratories (Jamaica), University of the West Indies, Kingston 7, Jamaica. Copyright © 1999 by Mosby, Inc. 0022-3476/99/$8.00 + 0 9/21/95473
304
bloodstream promotes an effective antibody response. Splenectomy in otherwise normal individuals increases the risk of overwhelming infections,1,2 especially with Streptococcus pneumoniae. In homozygous sickle cell disease, splenic function is damaged by the abnormal red cells early in life,3 rendering patients prone to overwhelming septicemia. Splenectomy may prevent deaths from recurrent acute splenic sequestration4-6 and relieve the hematologic sequelae of chronic hypersplenism.7-9 However, this operation may be deferred in favor of chronic
transfusion10,11 because of reluctance to remove remaining splenic immune function and the possibility of restoring lost function.12 This study examines the postoperative performance of 130 patients with SS disease who underwent splenectomy to determine whether this procedure increases the risk of serious complications. ASS CHS OR SS
Acute splenic sequestration Chronic hypersplenism Odds ratio Homozygous sickle cell disease
METHODS Patients The patients attended the Sickle Cell Clinic of the University Hospital of the West Indies, Kingston, Jamaica, and included 38 from the Jamaican Cohort Study based on newborn diagnosis and 92 from the general clinic, which serves predominantly patients referred with symptoms. The study was confined to 130 patients with SS disease undergoing splenectomy in a 22.5-year period (July 1, 1974 to December 31, 1996). The diagnosis of SS disease was based on standard methods.13 A nonsplenectomized control group was identified retrospectively by selecting for each index case the patient with SS of the same sex, closest in date of birth, and under follow-up at the date of splenectomy in the index case. To reduce potential bias this control group was selected from the register of all patients including those who have since died, emigrated, or been lost to followup. Frequency of attendance and clinical events detected were likely to be in-
WRIGHT ET AL
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3 fluenced by the distance of residence from the clinic, but patients were not matched by this criterion, and analysis showed no difference between the rural and urban distribution between the control group and patients who underwent splenectomy for ASS (χ2 test = 0.82, df = 1, P = .37) or hypersplenism (χ2 test = 1.70, df = 1, P = .19). Two factors related to splenic persistence, HbF level and α+ thalassemia, when examined by conditional logistic regression showed no differences between the index and control groups for HbF level (n = 112, χ2 = 0.6, df = 1, P = .46) or α+ thalassemia (n = 56, χ2 = 2.46, df = 1, P = .15).
Protocol Clinical details were obtained by review of notes from the Sickle Cell Clinic, the University Hospital and, where necessary, other hospitals. Events recorded included immunization status (pneumococcal and Haemophilus influenzae type b), penicillin prophylaxis by monthly intramuscular injections, deaths, bacteremias, episodes of painful crisis, acute chest syndrome, and febrile episodes. Bacteremias and deaths were recorded over the entire postsplenectomy period up to December 31, 1996. Painful crises, acute chest syndrome, and febrile episodes after splenectomy were analyzed in the 90 pairs with 5 years of postsplenectomy observation but excluding the first 2 postoperative weeks, when events may have been attributable to anesthetic or surgical complications. Follow-up of all subjects was continued up to December 31, 1996, or until death or emigration except in 2 defaulted patients, 1 member of the case group and 1 member of the control group.
Definitions Bacteremia was defined as a febrile illness associated with a pathogen on blood culture. The painful crisis was defined as bone pain of sufficient severity to require parenteral narcotic
analgesia; episodes within 14 days were arbitrarily considered the same event. Acute chest syndrome was defined by radiologic confirmation of a new pulmonary infiltrate or signs of consolidation and was usually associated with a history of cough and dyspnea. Febrile episodes were defined as fever >101.4° F (38.5° C) without other identifiable cause in which microbiologic cultures of blood, urine, and stool were negative. ASS was defined as a sudden increase in splenic size (usually ≥3 cm below the costal margin), a fall in hemoglobin (by >2 g/dL and usually <4.5 g/dL), and an increase in reticulocyte count (usually 2 to 3 times steady state values), with a decrease in splenomegaly after transfusion or spontaneous resolution of the attack. Hypersplenism was defined as the combination of splenomegaly (usually ≥4 and often >6 cm below the costal margin), low hemoglobin (usually <6 g/dL), and high reticulocytes (usually >20%) sustained for periods greater than 3 months.
Indications for Splenectomy Splenectomy was performed for the prevention of recurrent ASS in 46 (35%) patients, usually after 2 attacks, and for relief of hypersplenism in 84 (65%) patients, usually after an observation period of 6 months.
Infection Prophylaxis Pneumococcal prophylaxis varied over the 22.5-year observation period but was always based on monthly injections of slow release penicillin and pneumococcal vaccine. A trial of penicillin prophylaxis in 1978 used penicillin in patients from 6 months to 3 years of age and the vaccine in patients 1 to 2 years of age. After that trial was published,14 penicillin was continued to 4 years, and the vaccine was given with the last dose of penicillin. Most patients undergoing splenectomy received pneumococcal and hemophilus influenzae type b vaccine before they underwent splenectomy, and they re-
ceived penicillin for 3 years after surgery.
Follow-up Members of the index case group and the control group were closely matched for duration of follow-up (Log rank test, χ2 = 0.82, df = 1, P = .79), but follow-up in the ASS group (median 8.5 years, interquartile range 3.4 to 13.7 years) exceeded that in the control group (median 7.4 years, interquartile range 1.8 to 13.0 years), whereas in the CHS group (median 6.5 years, interquartile range 3.6 to 13.9 years) it was less than in the control group (median 7.2 years, interquartile range 3.6 to 12.5 years). The total patient-years of observation in these 4 groups were 389, 360, 711, and 690, respectively. Analysis of postsplenectomy events (painful crisis, acute chest syndrome, febrile episodes) was confined to those completing 5 years of follow-up, which excluded 17 (37%) members of the ASS group and 23 (27%) members of the CHS group, leaving 90 members of the case group (29 ASS, 61 CHS) and 90 members of the associated control group.
Statistical Methods Age at splenectomy in patients in the ASS and CHS groups was compared by the log-rank test. Domicile location, HbF levels, and α-thalassemia were compared between index cases and the control group by conditional logistic regression on the matched data set. Pneumococcal and Haemophilus influenzae type b vaccine use was compared between the splenectomy and control groups by the χ2 test. Duration of penicillin use was compared with the log-rank test. Mortality and bacteremia were compared with Fisher’s exact test, and time to death and bacteremia were examined by the log-rank test. The effect of time, the case/control grouping, and the splenectomy indication (ASS/CHS) on the annual frequency of 3 postsplenectomy clinical outcomes (painful crisis, acute chest 305
WRIGHT ET AL
THE JOURNAL OF PEDIATRICS MARCH 1999
Table I. Deaths in observation period in splenectomized group and control group (splenectomy group ranked by age at splenectomy, control group by age at death)
Splenectomy Patient No. Splenectomy group 1S 2S 3S 4S 5S 6S 7S 8S 9S Control group 1C 2C 3C 4C 5C 6C 7C 8C 9C 10C 11C 12C
Reason
Age
Age at death
ASS ASS ASS CHS CHS CHS CHS CHS CHS
1.3 2.9 3.1 4.6 4.9 5.4 10.9 10.9 13.8
6.8 10.1 3.8 7.1 5.0 13.8 11.3 11.2 27.4
Aplastic crisis Cerebrovascular accident Acute chest syndrome Unknown (PM not performed) No information available Acute chest syndrome Gastroenteritis, pulmonary edema Fever, acute hepatic sequestration Salmonella enteritidis septicemia
— — — — — — — — — — — —
— — — — — — — — — — — —
3.8 4.7 5.3 5.3 5.9 7.5 7.6 16.3 20.2 20.8 24.1 24.6
ASS superimposed on CHS Acute chest syndrome Staphylococcus aureus meningitis Salmonella enteritidis septicemia No information Acute chest syndrome, pulmonary edema Acute chest syndrome Cardiac failure (mitral valve disease) Pulmonary embolus Acute chest syndrome Staphylococcus aureus septicemia Road traffic accident
syndrome, febrile episodes), which were assumed to be independent, were explored by log linear modeling, after adjustment was done for the confounding effects of age at event and patient sex. Age was modeled in 5 groups (0 to 4 years, 5 to 9 years, 10 to 14 years, 15 to 19 years, 20+ years) where possible and in fewer groups if only sparse data were available. Statistical significance was assumed at the 95% level. All analyses were conducted with Stata 5 statistical software.15
RESULTS Of the 130 patients treated by splenectomy, 64 (49%) were male, and the 46 patients with splenectomy for ASS were younger at the time of 306
splenectomy (median age 2.3 years, interquartile range 1.5 to 3.4 years) than the 84 patients with CHS (5.9 years, interquartile range 3.7 to 10.8 years) (log rank test, χ2 = 34.1, df = 1, P < .001). Prophylaxis against infection continued for longer in the splenectomy group. Pneumococcal vaccine had been given before splenectomy in 112 (86%) members of the splenectomy group and in 64 (49%) members of the control group (χ2 = 40.5, df = 1, P < .001). Immunization against Haemophilus influenzae type b was given in 46 (35%) members of the splenectomy group and in 10 (8%) members of the control group (χ2 = 29.5, df = 1, P < .001). Penicillin prophylaxis had been given in 112 (86%) members of the postsplenectomy group compared with 62 (48%) members of the control group (χ2 =
Cause of death
43.4, df = 1, P < .001). In those receiving penicillin the duration was greater in the splenectomized group (median 36.6 months, interquartile range 24.3 to 62.3 months) than in the control group (median 27.5 months, interquartile range 19.1 to 31.0 months) (log-rank test χ2 = 18.7, df = 1, P < .001). Mortality (Table I) was not greater in the splenectomy group (Fisher’s exact test, P = .65). There were 9 deaths (1 with bacterial infection) compared with 12 in the control group (3 with bacterial infection), and no difference was seen between the groups in the interval between recruitment (splenectomy in the index case, same date in the control group) and death (log-rank test χ2 = 0.48, df = 1, P = .49). One of the deaths in the control group resulted from acute splenic sequestration super-
WRIGHT ET AL
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3
Table II. Bacteremia in splenectomy and control groups (splenectomy group ranked by age at splenectomy, control group by age at bacteremia)
Splenectomy Patient No. Splenectomy group 1S 2S 3S 4S 5S 6S 7S 8S (9S in Table I) 9S 10S Control group 1C 2C 3C 4C 5C (3C in Table I) 6C (4C in Table I) 7C 8C 9C 10C 11C 12C (11C in Table I)
Reason
Age
Age at bacteremia
CHS CHS ASS ASS ASS ASS ASS CHS CHS ASS
1.1 1.4 2.1 2.5 2.8 3.9 4.4 13.8 20.8 25.4
1.6 7.5 13.1 5.0 6.9 7.4 5.4 27.4 27.6 29.7
Streptococcus pneumoniae Streptococcus pneumoniae Clostridium welchii Streptococcus pneumoniae Enterobacter Salmonella enteritidis Haemophilus influenzae type b Salmonella enteritidis (died) Streptococcus pneumoniae Escherichia coli
— — — — — — — — — — — —
— — — — — — — — — — — —
1.5 2.4 2.5 4.9 5.3 5.3 5.4 5.7 7.0 13.7 14.3 24.1
Haemophilus influenzae type b Bacillus spp Salmonella spp. Enterobacter spp Staphylococcus aureus (died) Salmonella enteritidis (died) Escherichia coli Pseudomonas spp Salmonella enteritidis Streptococcus pneumoniae Salmonella spp. Staphylococcus aureus (died)
imposed on hypersplenism while the patient was awaiting splenectomy. Bacteremia (Table II) was not more prevalent in the splenectomy group (Fisher’s exact test, P = .82). There were 10 bacteremias compared with 12 in the control group, with no difference between the groups in the interval between recruitment (splenectomy in the index case, same date in the control group) and bacteremia (log rank test χ2 = 0.32, df = 1, P = .57). Death from septicemia occurred in 1 patient who underwent splenectomy (Salmonella enteritidis) compared with 3 members of the control group (2 Staphylococcus aureus, 1 Salmonella enteritidis). Streptococcus pneumoniae was not associated with death, but pneumococcal bacteremias occurred in 4 patients who underwent splenectomy and in 1 member of the control group. Bacteremia
with Salmonella species occurred in 2 patients who underwent splenectomy (1 death) and in 4 controls (1 death). The remaining bacteremias were predominantly caused by gram negative organisms. Painful crises were more common among patients who underwent splenectomy (odds ratio = 1.47 [95% confidence intervals 1.10, 1.98], P = .01), but the risk did not change with time over the 5-year period after splenectomy (OR = 1.03 [0.93, 1.14], P = .57) and did not differ between splenectomy indications (OR = 1.31 [0.83, 2.08], P = .25). Acute chest syndrome was more common among patients who underwent splenectomy (OR = 1.88 [1.25, 2.84], P = .002), but the risk did not change with time over the 5-year period after splenectomy (OR = 0.91 [0.79, 1.05], P = .18), al-
Organism
though it was greater for the ASS group than the CHS group (OR = 3.38 [1.53, 7.43], P = .01). Febrile episodes did not differ between the case and control groups (OR = 1.23 [0.59, 2.56], P = .58), between splenectomy indications (OR = 0.92 [0.71, 1.19], P = .51), or over the 5-year observation period (OR = 0.86 [0.29, 2.55], P = .78).
DISCUSSION The decision to perform splenectomy must be based on the relative risks and benefits of the procedure. The established risks involve those of general anesthesia, a surgical procedure, and increased postsplenectomy infections. Elective splenectomy may be performed safely with minimal postoperative morbidity and mortality,8,9 and the 307
WRIGHT ET AL
principal argument revolves around whether splenectomy in patients with SS disease further compromises their already impaired immune status. The risks of septicemia, particularly with pneumococci, are well recognized1,2 in otherwise normal patients undergoing splenectomy and have been reduced but not abolished by immunization and penicillin prophylaxis. In SS disease the spleen is damaged from early in life, and impaired splenic function is already manifest in an increased susceptibility to infection with encapsulated organisms. The principal indications for splenectomy in SS disease are the prevention of recurrent acute splenic sequestration and the relief of hypersplenism. Physicians managing these complications are aware that the natural history of SS disease is for a progressive splenic fibrosis and atrophy, and they tend to favor conservative measures while awaiting this outcome. Reluctance to advocate splenectomy stems from the theoretical risks of further compromising immune function, yet the spleen may contribute to morbidity and mortality. Acute splenic sequestration is a major cause of early death,4-6 and although this mortality has been reduced by parental education,6 survivors of a first attack are prone to recurrence and further mortality.5,6 Chronic transfusion may restore or improve splenic function,12 but it is unclear how to monitor its effect or when to stop such a program. Cessation of transfusion after 5 to 6 years, when ASS becomes rare in patients who do not undergo transfusion, may be followed by recurrent attacks,10 and even aggressive transfusion programs with HbS levels <20% may fail to prevent recurrence.11 Furthermore all chronic transfusion programs carry dangers and difficulties16 including the risks of erythrocyte alloimmunization, iron overload, transmission of bloodborne infections, increasing difficulties of venous access, and compliance. Splenectomy prevents 308
THE JOURNAL OF PEDIATRICS MARCH 1999 further ASS without the risks of chronic transfusion. Chronic hypersplenism causes excessive hemolysis with markedly shortened erythrocyte survival,7 low hemoglobin, high reticulocyte counts, and high energy costs, which may compromise growth and development. Death may occur from superimposed acute sequestration (1 patient in the control group died while awaiting splenectomy), coincidental aplastic crisis, or hemorrhage. Chronic transfusion programs have been advocated for hypersplenism, but the same reservations remain. Splenectomy is followed by a rapid improvement in blood counts and well-being with reduction in whole body protein turnover17 and an increase in linear growth.18 The data presented here confirm earlier impressions8,9 that splenectomy in SS disease is a safe procedure with low perioperative and long-term morbidity. Deaths and bacteremic episodes were no greater in the splenectomy group than in the control group. The extent to which this is due to the more vigilant prophylaxis after splenectomy cannot be ascertained, and it is possible that a potential decrease in immune function after splenectomy is compensated for by a higher frequency of immunization and more prolonged penicillin prophylaxis. Nevertheless, this clinical experience remains valid in reaching a decision on splenectomy. The high frequency of salmonella bacteremia (2 members of the splenectomy group, 4 members of the control group) reflects the increasing importance of this organism in SS disease in Jamaica.19 Splenectomy for chronic hypersplenism is usually followed by a striking rise in hemoglobin and platelet count,7 and because a high hemoglobin level is a risk factor for painful crises,20,21 an increased frequency of pain might have been anticipated in this group. Painful crises were more common in the patients who underwent splenectomy, but contrary to expectation, this difference occurred in
both the ASS and CHS groups. This increased prevalence remains unexplained and was not accounted for by known determinants of bone pain such as high hemoglobin, low fetal hemoglobin, or α-thalassemia, and it is assumed that factors still unrecognized may influence both splenic disease and frequency of bone pain crisis. Acute chest syndrome is a major cause of mortality22 and of postoperative morbidity.23 It was more common among patients who underwent splenectomy and among the ASS group, contrary to expectation, because the CHS group would be expected to manifest a rise in hemoglobin level, which may be a risk factor for the acute chest syndrome.24 A possible explanation could be that the members of the ASS group had intrinsically more severe sickle cell disease and hence were more prone to the acute chest syndrome. Another major difference between the groups was the younger age at splenectomy for ASS, and it remains possible that earlier splenectomy increases the risks of these complications. Febrile episodes did not differ between the splenectomy and control groups or between the indications for splenectomy. Splenectomy should not be deferred for fear of increasing serious morbidity or mortality. It should be used when necessary in the management of acute splenic sequestration or hypersplenism in SS disease.
REFERENCES 1. King H, Schumacher HB Jr. Splenic studies. Susceptibility to infection after splenectomy performed in infancy. Ann Surg 1952;136:239-42. 2. Schwartz PE, Sterioff S, Mucha P, Offord KP. Post splenectomy sepsis and mortality in adults. JAMA 1982;248: 2279-83. 3. Pearson HA, Gallagher D, Chilcote R, Sullivan E, Wilimas J, Espeland M, et al. Developmental pattern of splenic dysfunction in sickle cell disorders. Pediatrics 1985;76:392-7. 4. Seeler RA, Shwiaki M. Acute splenic
WRIGHT ET AL
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3
5.
6.
7.
8.
9.
10.
11.
sequestration crisis in young children with sickle cell anemia. Clin Pediatr 1972;11:701-4. Topley JM, Rogers DW, Stevens MC, Serjeant GR. Acute splenic sequestration and hypersplenism in the first 5 years of life in homozygous sickle cell disease. Arch Dis Child 1981;56:765-9. Emond AM, Collins R, Darvill D, Maude GH, Serjeant GR. Acute splenic sequestration in homozygous sickle cell disease: natural history and management. J Pediatr 1985;107:201-6. Rossi EC, Westring DW, Santos AJ, Gutierrez J. Hypersplenism in sickle cell anemia. Arch Intern Med 1964; 114;408-12. Emond AM, Morais P, Venugopal S, Morais P, Carpenter RG, Serjeant GR. Role of splenectomy in homozygous sickle cell disease in childhood. Lancet 1984;1:88-90. Al-Salem AH, Qaisaruddin S, Nasserallah Z, Al-Dabbous I, Jam’a AA. Splenectomy in patients with sickle cell disease. Am J Surg 1996;171:254-8. Rao S, Pang E. Transfusion therapy for subacute splenic sequestration in sickle cell disease. Blood 1982;60 (suppl):48a. Kinney TR, Ware RE, Schultz WH, Filston HC. Long-term management
12.
13.
14.
15.
16.
17.
18.
of splenic sequestration in children with sickle cell disease. J Pediatr 1990;117:194-9. Pearson HA, Cornelius EA, Schwartz AD, Zelson JH, Wolfson SL, Spencer RP. Transfusion reversible asplenia in young children with sickle-cell anemia. N Engl J Med 1970;283:334-7. Serjeant GR. Sickle Cell Disease. 2nd ed. Oxford Medical publication: Oxford University Press; 1992. John AB, Ramlal A, Jackson H, Maude GH, Waight-Sharma A, Serjeant GR. Prevention of pneumococcal infection in children with homozygous sickle cell disease. Br Med J 1984; 288:1567-70. Stata Corp. Stata Statistical Software: Release 5.0, 1997. College Station, Texas, Stata Corporation. Serjeant GR. Chronic transfusion regimes in sickle cell disease: problem or panacea? Br J Haematol 1997;97:253-5. Badaloo A, Emond A, Venugopal S, Serjeant GR, Jackson A. The effect of splenectomy on whole body protein turnover in homozygous sickle cell disease. Acta Paed Scand 1991;80:103-5. Singhal A, Thomas PW, Kearney T, Venugopal S, Serjeant GR. Acceleration in linear growth after splenectomy for hypersplenism in homozygous sick-
19.
20.
21.
22.
23.
24.
le cell disease. Arch Dis Child 1995; 72L:227-9. Wright J, Thomas P, Serjeant GR. Septicemia caused by salmonella infection; an overlooked complication of sickle cell disease. J Pediatr 1997;130: 394-9. Baum KF, Dunn DT, Maude GH, Serjeant GR. The painful crisis of homozygous sickle cell disease: a study of risk factors. Arch Intern Med 1987; 147:1231-4. Platt OS, Thorington BD, Brambilla DJ, Milner PF, Rosse WF, Vichinsky E, Kinney TR. Pain in sickle cell disease. Rates and risk factors. N Engl J Med 1991;325:11-6. Platt OS, Brambilla DJ, Rosse WF, Milner PF, Castro O, Steinberg MH, Klug PP. Mortality in sickle cell disease: life expectancy and risk factors for early death. N Engl J Med 1994; 330:1639-44. Homi J, Reynolds J, Skinner A, Hanna W, Serjeant GR. General anaesthesia in sickle cell disease. Br Med J 1979;1:1599-601. Castro O, Brambilla DJ, Thorington B, Reindorf CA, Scott RB, Gillette P, et al. The acute chest syndrome in sickle cell disease: incidence and risk factors. Blood 1994;84:643-9.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the 1999 issues of The Journal of Pediatrics are available to subscribers (only) from the Publisher, at a cost of $100.00 for domestic, $125.19 for Canadian, and $117.00 for international subscribers, for Vol. 134 (January-June) and Vol. 135 (July-December), shipping charges included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram, with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, Inc., Subscription Services, 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, USA/800-4534351, or 314-453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.
309
Postsplenectomy course in homozygous sickle cell disease
J. G. Wright, MB, MRCP, MRCPath, I. R. Hambleton, BA, MSc, P. W. Thomas, PhD, N. D. Duncan, DM, S. Venugopal, MS, FACS, FICS, and G. R. Serjeant, MD, FRCP
Objective: To determine whether children with homozygous sickle cell (SS) disease and splenectomy are at greater risk of death, overwhelming septicemia, or other complications. Methods: A total of 130 patients with SS treated by splenectomy (46 recurrent acute splenic sequestration, 84 chronic hypersplenism) over a 22.5-year period at the Sickle Cell Clinic of the University Hospital of the West Indies, Kingston, Jamaica, were compared with a control group matched for sex, age, and duration of follow-up in a retrospective review. Deaths and bacteremias were examined over the whole study period. Painful crises, acute chest syndromes, and febrile episodes were compared in the 90 patients completing 5 years of postsplenectomy follow-up. Findings: Mortality and bacteremic episodes did not differ between the splenectomy and control groups. Painful crises were more common in the splenectomy group than in the control group (P = .01) but did not differ between splenectomy indications. Acute chest syndrome was more common in the splenectomy group than in the control group (P < .01) and was more common in the acute splenic sequestration group than in the hypersplenism group (P = .01). Febrile events did not differ between the groups or between the indications for splenectomy. Conclusion: Splenectomy does not increase the risk of death or bacteremic illness in patients with SS disease and, if otherwise indicated, should not be deferred for these reasons. (J Pediatr 1999;134:304-9)
The complex vasculature of the spleen provides an efficient filter against bacteria, and the mass of reticuloendothelial cells in close proximity to the From the MRC Laboratories (Jamaica) and the Division of Paediatric Surgery, University of the West Indies, Kingston, Jamaica.
Submitted for publication May 8, 1998; revision received Sept 18, 1998; accepted Oct 28, 1998. Reprint requests: GR Serjeant, MRC Laboratories (Jamaica), University of the West Indies, Kingston 7, Jamaica. Copyright © 1999 by Mosby, Inc. 0022-3476/99/$8.00 + 0 9/21/95473
304
bloodstream promotes an effective antibody response. Splenectomy in otherwise normal individuals increases the risk of overwhelming infections,1,2 especially with Streptococcus pneumoniae. In homozygous sickle cell disease, splenic function is damaged by the abnormal red cells early in life,3 rendering patients prone to overwhelming septicemia. Splenectomy may prevent deaths from recurrent acute splenic sequestration4-6 and relieve the hematologic sequelae of chronic hypersplenism.7-9 However, this operation may be deferred in favor of chronic
transfusion10,11 because of reluctance to remove remaining splenic immune function and the possibility of restoring lost function.12 This study examines the postoperative performance of 130 patients with SS disease who underwent splenectomy to determine whether this procedure increases the risk of serious complications. ASS CHS OR SS
Acute splenic sequestration Chronic hypersplenism Odds ratio Homozygous sickle cell disease
METHODS Patients The patients attended the Sickle Cell Clinic of the University Hospital of the West Indies, Kingston, Jamaica, and included 38 from the Jamaican Cohort Study based on newborn diagnosis and 92 from the general clinic, which serves predominantly patients referred with symptoms. The study was confined to 130 patients with SS disease undergoing splenectomy in a 22.5-year period (July 1, 1974 to December 31, 1996). The diagnosis of SS disease was based on standard methods.13 A nonsplenectomized control group was identified retrospectively by selecting for each index case the patient with SS of the same sex, closest in date of birth, and under follow-up at the date of splenectomy in the index case. To reduce potential bias this control group was selected from the register of all patients including those who have since died, emigrated, or been lost to followup. Frequency of attendance and clinical events detected were likely to be in-
WRIGHT ET AL
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3 fluenced by the distance of residence from the clinic, but patients were not matched by this criterion, and analysis showed no difference between the rural and urban distribution between the control group and patients who underwent splenectomy for ASS (χ2 test = 0.82, df = 1, P = .37) or hypersplenism (χ2 test = 1.70, df = 1, P = .19). Two factors related to splenic persistence, HbF level and α+ thalassemia, when examined by conditional logistic regression showed no differences between the index and control groups for HbF level (n = 112, χ2 = 0.6, df = 1, P = .46) or α+ thalassemia (n = 56, χ2 = 2.46, df = 1, P = .15).
Protocol Clinical details were obtained by review of notes from the Sickle Cell Clinic, the University Hospital and, where necessary, other hospitals. Events recorded included immunization status (pneumococcal and Haemophilus influenzae type b), penicillin prophylaxis by monthly intramuscular injections, deaths, bacteremias, episodes of painful crisis, acute chest syndrome, and febrile episodes. Bacteremias and deaths were recorded over the entire postsplenectomy period up to December 31, 1996. Painful crises, acute chest syndrome, and febrile episodes after splenectomy were analyzed in the 90 pairs with 5 years of postsplenectomy observation but excluding the first 2 postoperative weeks, when events may have been attributable to anesthetic or surgical complications. Follow-up of all subjects was continued up to December 31, 1996, or until death or emigration except in 2 defaulted patients, 1 member of the case group and 1 member of the control group.
Definitions Bacteremia was defined as a febrile illness associated with a pathogen on blood culture. The painful crisis was defined as bone pain of sufficient severity to require parenteral narcotic
analgesia; episodes within 14 days were arbitrarily considered the same event. Acute chest syndrome was defined by radiologic confirmation of a new pulmonary infiltrate or signs of consolidation and was usually associated with a history of cough and dyspnea. Febrile episodes were defined as fever >101.4° F (38.5° C) without other identifiable cause in which microbiologic cultures of blood, urine, and stool were negative. ASS was defined as a sudden increase in splenic size (usually ≥3 cm below the costal margin), a fall in hemoglobin (by >2 g/dL and usually <4.5 g/dL), and an increase in reticulocyte count (usually 2 to 3 times steady state values), with a decrease in splenomegaly after transfusion or spontaneous resolution of the attack. Hypersplenism was defined as the combination of splenomegaly (usually ≥4 and often >6 cm below the costal margin), low hemoglobin (usually <6 g/dL), and high reticulocytes (usually >20%) sustained for periods greater than 3 months.
Indications for Splenectomy Splenectomy was performed for the prevention of recurrent ASS in 46 (35%) patients, usually after 2 attacks, and for relief of hypersplenism in 84 (65%) patients, usually after an observation period of 6 months.
Infection Prophylaxis Pneumococcal prophylaxis varied over the 22.5-year observation period but was always based on monthly injections of slow release penicillin and pneumococcal vaccine. A trial of penicillin prophylaxis in 1978 used penicillin in patients from 6 months to 3 years of age and the vaccine in patients 1 to 2 years of age. After that trial was published,14 penicillin was continued to 4 years, and the vaccine was given with the last dose of penicillin. Most patients undergoing splenectomy received pneumococcal and hemophilus influenzae type b vaccine before they underwent splenectomy, and they re-
ceived penicillin for 3 years after surgery.
Follow-up Members of the index case group and the control group were closely matched for duration of follow-up (Log rank test, χ2 = 0.82, df = 1, P = .79), but follow-up in the ASS group (median 8.5 years, interquartile range 3.4 to 13.7 years) exceeded that in the control group (median 7.4 years, interquartile range 1.8 to 13.0 years), whereas in the CHS group (median 6.5 years, interquartile range 3.6 to 13.9 years) it was less than in the control group (median 7.2 years, interquartile range 3.6 to 12.5 years). The total patient-years of observation in these 4 groups were 389, 360, 711, and 690, respectively. Analysis of postsplenectomy events (painful crisis, acute chest syndrome, febrile episodes) was confined to those completing 5 years of follow-up, which excluded 17 (37%) members of the ASS group and 23 (27%) members of the CHS group, leaving 90 members of the case group (29 ASS, 61 CHS) and 90 members of the associated control group.
Statistical Methods Age at splenectomy in patients in the ASS and CHS groups was compared by the log-rank test. Domicile location, HbF levels, and α-thalassemia were compared between index cases and the control group by conditional logistic regression on the matched data set. Pneumococcal and Haemophilus influenzae type b vaccine use was compared between the splenectomy and control groups by the χ2 test. Duration of penicillin use was compared with the log-rank test. Mortality and bacteremia were compared with Fisher’s exact test, and time to death and bacteremia were examined by the log-rank test. The effect of time, the case/control grouping, and the splenectomy indication (ASS/CHS) on the annual frequency of 3 postsplenectomy clinical outcomes (painful crisis, acute chest 305
WRIGHT ET AL
THE JOURNAL OF PEDIATRICS MARCH 1999
Table I. Deaths in observation period in splenectomized group and control group (splenectomy group ranked by age at splenectomy, control group by age at death)
Splenectomy Patient No. Splenectomy group 1S 2S 3S 4S 5S 6S 7S 8S 9S Control group 1C 2C 3C 4C 5C 6C 7C 8C 9C 10C 11C 12C
Reason
Age
Age at death
ASS ASS ASS CHS CHS CHS CHS CHS CHS
1.3 2.9 3.1 4.6 4.9 5.4 10.9 10.9 13.8
6.8 10.1 3.8 7.1 5.0 13.8 11.3 11.2 27.4
Aplastic crisis Cerebrovascular accident Acute chest syndrome Unknown (PM not performed) No information available Acute chest syndrome Gastroenteritis, pulmonary edema Fever, acute hepatic sequestration Salmonella enteritidis septicemia
— — — — — — — — — — — —
— — — — — — — — — — — —
3.8 4.7 5.3 5.3 5.9 7.5 7.6 16.3 20.2 20.8 24.1 24.6
ASS superimposed on CHS Acute chest syndrome Staphylococcus aureus meningitis Salmonella enteritidis septicemia No information Acute chest syndrome, pulmonary edema Acute chest syndrome Cardiac failure (mitral valve disease) Pulmonary embolus Acute chest syndrome Staphylococcus aureus septicemia Road traffic accident
syndrome, febrile episodes), which were assumed to be independent, were explored by log linear modeling, after adjustment was done for the confounding effects of age at event and patient sex. Age was modeled in 5 groups (0 to 4 years, 5 to 9 years, 10 to 14 years, 15 to 19 years, 20+ years) where possible and in fewer groups if only sparse data were available. Statistical significance was assumed at the 95% level. All analyses were conducted with Stata 5 statistical software.15
RESULTS Of the 130 patients treated by splenectomy, 64 (49%) were male, and the 46 patients with splenectomy for ASS were younger at the time of 306
splenectomy (median age 2.3 years, interquartile range 1.5 to 3.4 years) than the 84 patients with CHS (5.9 years, interquartile range 3.7 to 10.8 years) (log rank test, χ2 = 34.1, df = 1, P < .001). Prophylaxis against infection continued for longer in the splenectomy group. Pneumococcal vaccine had been given before splenectomy in 112 (86%) members of the splenectomy group and in 64 (49%) members of the control group (χ2 = 40.5, df = 1, P < .001). Immunization against Haemophilus influenzae type b was given in 46 (35%) members of the splenectomy group and in 10 (8%) members of the control group (χ2 = 29.5, df = 1, P < .001). Penicillin prophylaxis had been given in 112 (86%) members of the postsplenectomy group compared with 62 (48%) members of the control group (χ2 =
Cause of death
43.4, df = 1, P < .001). In those receiving penicillin the duration was greater in the splenectomized group (median 36.6 months, interquartile range 24.3 to 62.3 months) than in the control group (median 27.5 months, interquartile range 19.1 to 31.0 months) (log-rank test χ2 = 18.7, df = 1, P < .001). Mortality (Table I) was not greater in the splenectomy group (Fisher’s exact test, P = .65). There were 9 deaths (1 with bacterial infection) compared with 12 in the control group (3 with bacterial infection), and no difference was seen between the groups in the interval between recruitment (splenectomy in the index case, same date in the control group) and death (log-rank test χ2 = 0.48, df = 1, P = .49). One of the deaths in the control group resulted from acute splenic sequestration super-
WRIGHT ET AL
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3
Table II. Bacteremia in splenectomy and control groups (splenectomy group ranked by age at splenectomy, control group by age at bacteremia)
Splenectomy Patient No. Splenectomy group 1S 2S 3S 4S 5S 6S 7S 8S (9S in Table I) 9S 10S Control group 1C 2C 3C 4C 5C (3C in Table I) 6C (4C in Table I) 7C 8C 9C 10C 11C 12C (11C in Table I)
Reason
Age
Age at bacteremia
CHS CHS ASS ASS ASS ASS ASS CHS CHS ASS
1.1 1.4 2.1 2.5 2.8 3.9 4.4 13.8 20.8 25.4
1.6 7.5 13.1 5.0 6.9 7.4 5.4 27.4 27.6 29.7
Streptococcus pneumoniae Streptococcus pneumoniae Clostridium welchii Streptococcus pneumoniae Enterobacter Salmonella enteritidis Haemophilus influenzae type b Salmonella enteritidis (died) Streptococcus pneumoniae Escherichia coli
— — — — — — — — — — — —
— — — — — — — — — — — —
1.5 2.4 2.5 4.9 5.3 5.3 5.4 5.7 7.0 13.7 14.3 24.1
Haemophilus influenzae type b Bacillus spp Salmonella spp. Enterobacter spp Staphylococcus aureus (died) Salmonella enteritidis (died) Escherichia coli Pseudomonas spp Salmonella enteritidis Streptococcus pneumoniae Salmonella spp. Staphylococcus aureus (died)
imposed on hypersplenism while the patient was awaiting splenectomy. Bacteremia (Table II) was not more prevalent in the splenectomy group (Fisher’s exact test, P = .82). There were 10 bacteremias compared with 12 in the control group, with no difference between the groups in the interval between recruitment (splenectomy in the index case, same date in the control group) and bacteremia (log rank test χ2 = 0.32, df = 1, P = .57). Death from septicemia occurred in 1 patient who underwent splenectomy (Salmonella enteritidis) compared with 3 members of the control group (2 Staphylococcus aureus, 1 Salmonella enteritidis). Streptococcus pneumoniae was not associated with death, but pneumococcal bacteremias occurred in 4 patients who underwent splenectomy and in 1 member of the control group. Bacteremia
with Salmonella species occurred in 2 patients who underwent splenectomy (1 death) and in 4 controls (1 death). The remaining bacteremias were predominantly caused by gram negative organisms. Painful crises were more common among patients who underwent splenectomy (odds ratio = 1.47 [95% confidence intervals 1.10, 1.98], P = .01), but the risk did not change with time over the 5-year period after splenectomy (OR = 1.03 [0.93, 1.14], P = .57) and did not differ between splenectomy indications (OR = 1.31 [0.83, 2.08], P = .25). Acute chest syndrome was more common among patients who underwent splenectomy (OR = 1.88 [1.25, 2.84], P = .002), but the risk did not change with time over the 5-year period after splenectomy (OR = 0.91 [0.79, 1.05], P = .18), al-
Organism
though it was greater for the ASS group than the CHS group (OR = 3.38 [1.53, 7.43], P = .01). Febrile episodes did not differ between the case and control groups (OR = 1.23 [0.59, 2.56], P = .58), between splenectomy indications (OR = 0.92 [0.71, 1.19], P = .51), or over the 5-year observation period (OR = 0.86 [0.29, 2.55], P = .78).
DISCUSSION The decision to perform splenectomy must be based on the relative risks and benefits of the procedure. The established risks involve those of general anesthesia, a surgical procedure, and increased postsplenectomy infections. Elective splenectomy may be performed safely with minimal postoperative morbidity and mortality,8,9 and the 307
WRIGHT ET AL
principal argument revolves around whether splenectomy in patients with SS disease further compromises their already impaired immune status. The risks of septicemia, particularly with pneumococci, are well recognized1,2 in otherwise normal patients undergoing splenectomy and have been reduced but not abolished by immunization and penicillin prophylaxis. In SS disease the spleen is damaged from early in life, and impaired splenic function is already manifest in an increased susceptibility to infection with encapsulated organisms. The principal indications for splenectomy in SS disease are the prevention of recurrent acute splenic sequestration and the relief of hypersplenism. Physicians managing these complications are aware that the natural history of SS disease is for a progressive splenic fibrosis and atrophy, and they tend to favor conservative measures while awaiting this outcome. Reluctance to advocate splenectomy stems from the theoretical risks of further compromising immune function, yet the spleen may contribute to morbidity and mortality. Acute splenic sequestration is a major cause of early death,4-6 and although this mortality has been reduced by parental education,6 survivors of a first attack are prone to recurrence and further mortality.5,6 Chronic transfusion may restore or improve splenic function,12 but it is unclear how to monitor its effect or when to stop such a program. Cessation of transfusion after 5 to 6 years, when ASS becomes rare in patients who do not undergo transfusion, may be followed by recurrent attacks,10 and even aggressive transfusion programs with HbS levels <20% may fail to prevent recurrence.11 Furthermore all chronic transfusion programs carry dangers and difficulties16 including the risks of erythrocyte alloimmunization, iron overload, transmission of bloodborne infections, increasing difficulties of venous access, and compliance. Splenectomy prevents 308
THE JOURNAL OF PEDIATRICS MARCH 1999 further ASS without the risks of chronic transfusion. Chronic hypersplenism causes excessive hemolysis with markedly shortened erythrocyte survival,7 low hemoglobin, high reticulocyte counts, and high energy costs, which may compromise growth and development. Death may occur from superimposed acute sequestration (1 patient in the control group died while awaiting splenectomy), coincidental aplastic crisis, or hemorrhage. Chronic transfusion programs have been advocated for hypersplenism, but the same reservations remain. Splenectomy is followed by a rapid improvement in blood counts and well-being with reduction in whole body protein turnover17 and an increase in linear growth.18 The data presented here confirm earlier impressions8,9 that splenectomy in SS disease is a safe procedure with low perioperative and long-term morbidity. Deaths and bacteremic episodes were no greater in the splenectomy group than in the control group. The extent to which this is due to the more vigilant prophylaxis after splenectomy cannot be ascertained, and it is possible that a potential decrease in immune function after splenectomy is compensated for by a higher frequency of immunization and more prolonged penicillin prophylaxis. Nevertheless, this clinical experience remains valid in reaching a decision on splenectomy. The high frequency of salmonella bacteremia (2 members of the splenectomy group, 4 members of the control group) reflects the increasing importance of this organism in SS disease in Jamaica.19 Splenectomy for chronic hypersplenism is usually followed by a striking rise in hemoglobin and platelet count,7 and because a high hemoglobin level is a risk factor for painful crises,20,21 an increased frequency of pain might have been anticipated in this group. Painful crises were more common in the patients who underwent splenectomy, but contrary to expectation, this difference occurred in
both the ASS and CHS groups. This increased prevalence remains unexplained and was not accounted for by known determinants of bone pain such as high hemoglobin, low fetal hemoglobin, or α-thalassemia, and it is assumed that factors still unrecognized may influence both splenic disease and frequency of bone pain crisis. Acute chest syndrome is a major cause of mortality22 and of postoperative morbidity.23 It was more common among patients who underwent splenectomy and among the ASS group, contrary to expectation, because the CHS group would be expected to manifest a rise in hemoglobin level, which may be a risk factor for the acute chest syndrome.24 A possible explanation could be that the members of the ASS group had intrinsically more severe sickle cell disease and hence were more prone to the acute chest syndrome. Another major difference between the groups was the younger age at splenectomy for ASS, and it remains possible that earlier splenectomy increases the risks of these complications. Febrile episodes did not differ between the splenectomy and control groups or between the indications for splenectomy. Splenectomy should not be deferred for fear of increasing serious morbidity or mortality. It should be used when necessary in the management of acute splenic sequestration or hypersplenism in SS disease.
REFERENCES 1. King H, Schumacher HB Jr. Splenic studies. Susceptibility to infection after splenectomy performed in infancy. Ann Surg 1952;136:239-42. 2. Schwartz PE, Sterioff S, Mucha P, Offord KP. Post splenectomy sepsis and mortality in adults. JAMA 1982;248: 2279-83. 3. Pearson HA, Gallagher D, Chilcote R, Sullivan E, Wilimas J, Espeland M, et al. Developmental pattern of splenic dysfunction in sickle cell disorders. Pediatrics 1985;76:392-7. 4. Seeler RA, Shwiaki M. Acute splenic
WRIGHT ET AL
THE JOURNAL OF PEDIATRICS VOLUME 134, NUMBER 3
5.
6.
7.
8.
9.
10.
11.
sequestration crisis in young children with sickle cell anemia. Clin Pediatr 1972;11:701-4. Topley JM, Rogers DW, Stevens MC, Serjeant GR. Acute splenic sequestration and hypersplenism in the first 5 years of life in homozygous sickle cell disease. Arch Dis Child 1981;56:765-9. Emond AM, Collins R, Darvill D, Maude GH, Serjeant GR. Acute splenic sequestration in homozygous sickle cell disease: natural history and management. J Pediatr 1985;107:201-6. Rossi EC, Westring DW, Santos AJ, Gutierrez J. Hypersplenism in sickle cell anemia. Arch Intern Med 1964; 114;408-12. Emond AM, Morais P, Venugopal S, Morais P, Carpenter RG, Serjeant GR. Role of splenectomy in homozygous sickle cell disease in childhood. Lancet 1984;1:88-90. Al-Salem AH, Qaisaruddin S, Nasserallah Z, Al-Dabbous I, Jam’a AA. Splenectomy in patients with sickle cell disease. Am J Surg 1996;171:254-8. Rao S, Pang E. Transfusion therapy for subacute splenic sequestration in sickle cell disease. Blood 1982;60 (suppl):48a. Kinney TR, Ware RE, Schultz WH, Filston HC. Long-term management
12.
13.
14.
15.
16.
17.
18.
of splenic sequestration in children with sickle cell disease. J Pediatr 1990;117:194-9. Pearson HA, Cornelius EA, Schwartz AD, Zelson JH, Wolfson SL, Spencer RP. Transfusion reversible asplenia in young children with sickle-cell anemia. N Engl J Med 1970;283:334-7. Serjeant GR. Sickle Cell Disease. 2nd ed. Oxford Medical publication: Oxford University Press; 1992. John AB, Ramlal A, Jackson H, Maude GH, Waight-Sharma A, Serjeant GR. Prevention of pneumococcal infection in children with homozygous sickle cell disease. Br Med J 1984; 288:1567-70. Stata Corp. Stata Statistical Software: Release 5.0, 1997. College Station, Texas, Stata Corporation. Serjeant GR. Chronic transfusion regimes in sickle cell disease: problem or panacea? Br J Haematol 1997;97:253-5. Badaloo A, Emond A, Venugopal S, Serjeant GR, Jackson A. The effect of splenectomy on whole body protein turnover in homozygous sickle cell disease. Acta Paed Scand 1991;80:103-5. Singhal A, Thomas PW, Kearney T, Venugopal S, Serjeant GR. Acceleration in linear growth after splenectomy for hypersplenism in homozygous sick-
19.
20.
21.
22.
23.
24.
le cell disease. Arch Dis Child 1995; 72L:227-9. Wright J, Thomas P, Serjeant GR. Septicemia caused by salmonella infection; an overlooked complication of sickle cell disease. J Pediatr 1997;130: 394-9. Baum KF, Dunn DT, Maude GH, Serjeant GR. The painful crisis of homozygous sickle cell disease: a study of risk factors. Arch Intern Med 1987; 147:1231-4. Platt OS, Thorington BD, Brambilla DJ, Milner PF, Rosse WF, Vichinsky E, Kinney TR. Pain in sickle cell disease. Rates and risk factors. N Engl J Med 1991;325:11-6. Platt OS, Brambilla DJ, Rosse WF, Milner PF, Castro O, Steinberg MH, Klug PP. Mortality in sickle cell disease: life expectancy and risk factors for early death. N Engl J Med 1994; 330:1639-44. Homi J, Reynolds J, Skinner A, Hanna W, Serjeant GR. General anaesthesia in sickle cell disease. Br Med J 1979;1:1599-601. Castro O, Brambilla DJ, Thorington B, Reindorf CA, Scott RB, Gillette P, et al. The acute chest syndrome in sickle cell disease: incidence and risk factors. Blood 1994;84:643-9.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the 1999 issues of The Journal of Pediatrics are available to subscribers (only) from the Publisher, at a cost of $100.00 for domestic, $125.19 for Canadian, and $117.00 for international subscribers, for Vol. 134 (January-June) and Vol. 135 (July-December), shipping charges included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram, with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, Inc., Subscription Services, 11830 Westline Industrial Dr., St. Louis, MO 63146-3318, USA/800-4534351, or 314-453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.
309